The Voice Beneath The Feet

AAn elephant that stands motionless with its ears spread and one forefoot lifted clear of the ground may not be resting. Since 1986, when tape recordings made at a North American zoo were replayed at ten times their original speed, biologists have known that African elephants produce rumbles whose fundamental frequency descends to about 14 hertz, well below the lower limit of human hearing. Low frequencies shed energy slowly, and on a still evening such a call can reach another elephant some two and a half kilometres away. Less obviously, every rumble also sets the ground itself moving. Part of the acoustic energy couples into the soil and spreads outward as a surface wave at roughly 250 metres per second, arriving at a distant animal a fraction of a second after the airborne version and, in dry savanna, still measurable once the airborne version has faded. Whether elephants make any use of this second pathway is a separate question, and a contested one.

BPriya Aluwihare, a bioacoustician at the Kalapani Institute of Sensory Ecology, has assembled the strongest case that they do. Between 2011 and 2019 her team buried geophones at forty-two locations across the Nkasi Pan and delivered recorded alarm rumbles into the soil with a shaker plate, an instrument that generates a seismic wave without any accompanying sound in air. Of the 118 trials that met her criteria, elephants within 800 metres froze, turned toward the source or raised a forefoot in 89 per cent of cases; the same animals reacted to a seismically delivered rumble taken from a population they had never encountered in only 31 per cent. From this Aluwihare argued that the seismic pathway is not a by-product but a message in its own right: the animals do not merely feel a disturbance, she wrote, they register whose disturbance it is. The claim is not that ground vibration reaches elephants, which nobody disputes, but that it carries information they act upon.

CAnatomy has been recruited in support. The foot of an elephant contains dense clusters of Pacinian corpuscles, pressure receptors tuned to vibration in the range from 20 to 300 hertz, and the fatty cushion beneath the toes is thought to carry vibration up the leg to the shoulder and from there, by bone conduction, to the middle ear. A modelling study published in 2016 estimated that an animal standing on firm ground could in principle sense a rumble travelling seismically from as far as six kilometres, more than twice the distance reported for the airborne route. The estimate cuts both ways. It shows how much reach the pathway might offer, yet the same model found that detection collapses once the substrate is loose sand or waterlogged clay, so the advantage exists only where the ground cooperates.

DTomas Berglund, who studies sensory ecology at the Ostrand Laboratory of Comparative Physiology, accepts the anatomy and disputes what may be inferred from it. His first objection concerns the playbacks themselves: a shaker plate reproduces the waveform of a rumble but not its origin, and an elephant orienting toward an unexplained tremor is behaving sensibly whether or not it has understood anything at all. His second concerns the 2016 figure, which was derived from a model rather than from any animal, and which assumed a uniform substrate that he says exists nowhere in the Nkasi Pan. The third is the one he presses hardest: in nineteen of Aluwihare's own trials the wind was such that a faint airborne component cannot be excluded, and until a trial is run in which the airborne path is demonstrably absent, the seismic result explains nothing that the ordinary hearing of elephants does not already explain. He grants that his own recordings were gathered in a single wet season.

EAluwihare replies that a shaker plate is the only means of isolating the pathway at all, and that the distance between her two figures cannot be explained as startle, since a startled animal has no way of knowing whose rumble it has failed to recognise. On the question of wind she is more accommodating, noting that the nineteen affected trials were removed from a later reanalysis and that the contrast survived their removal. She concedes further that no experiment of hers has followed a single animal from detection through to any consequent action, that her response categories were scored from video by observers who knew which recording was being played, and that the six-kilometre figure, which is not hers, has been quoted in her support more often than she would like.

FThe disagreement has since acquired an audience outside the journals. Vibration from quarrying, heavy road traffic and exploration surveys occupies the same low band as elephant rumbles, and one reading of Aluwihare's work holds that a road driven through a corridor may sever communication between herds as effectively as a fence. Submissions opposing a haulage route across the southern edge of the Nkasi Pan quote her papers at length. Berglund regards that reasoning as plausible in principle and untested in detail, since nobody has measured what a working quarry does to the calling behaviour of a herd. Aluwihare, whose figures appear in those submissions, has said that the route should be opposed on grounds that do not depend on her hypothesis being correct.